kuberay icon indicating copy to clipboard operation
kuberay copied to clipboard

[Feature] [kubectl-plugin] Add clusterTimeout flag support for job submit command

Open martynas-subonis opened this issue 5 months ago • 3 comments

Search before asking

  • [x] I had searched in the issues and found no similar feature requirement.

Description

Hello kuberay contributors đź‘‹

I'd like to propose a minor feature for the kubectl plugin, which is useful when using RayJob flow with submissionMode: InteractiveMode.

Currently, the clusterTimeout is hardcoded to 120 seconds. When submitting RayJobs as:

kubectl ray job submit -f ray_job_manifest.yaml

This becomes problematic when RayJob triggers node pool scaling events—especially on cloud providers where scaling can take several minutes (~10 min in some cases). The hardcoded timeout causes premature failures, forcing workarounds on the client side.

It would be valuable to make this timeout configurable, allowing users to adapt it to their cluster characteristics.

Thanks for considering

Use case

No response

Related issues

No response

Are you willing to submit a PR?

  • [x] Yes I am willing to submit a PR!

martynas-subonis avatar Jul 25 '25 11:07 martynas-subonis

I have observed the same issue, would be great if this could be addressed soon.

nkaenzig avatar Aug 12 '25 16:08 nkaenzig

Woud be great if this could be added!

daankaiko avatar Aug 13 '25 06:08 daankaiko

why don't use ctrl+c to delete your CR?

Future-Outlier avatar Oct 11 '25 01:10 Future-Outlier